Forgetfulness is one of the most important symptoms in many neurocognitive diseases and especially known from Alzheimer’s disease. Forgetting is generally considered to be a flaw in our memory function. In fact, it appears to reflect the structure of our memory function and even has an adaptive component. We commonly understand forgetting as the deletion or non-existence of a memory trace. However, Shiffrin (1970) already said that forgetting might be a problem in retrieving the memory trace without it necessarily having been destroyed beforehand.
The fact that memory traces may not be completely eliminated is also evident in Loftus’ (1977) false information paradigm. In this experiment, subjects are first presented with pictures of an accident between a pedestrian and a car, while a green car passes by the scene. After the pictures were presented, the subjects were asked to answer questions. The control group was asked “Did the car that passed the accident have a ski rack on its roof?” while the experimental group was presented with incorrect information in their question: “Did the blue car that passed the accident have a ski rack on its roof?”. When the subjects were asked afterwards to indicate the colour of the car, there was a shift in colour towards blue in the false-information group. However, the colour was not a complete blue but a mixed colour of blue and green, which contradicts a deletion of the old trace. Instead, the results tended to indicate some kind of interference between the two traces1. In fact, there is further evidence for the interference of different memory traces.
Typically, memory content is associated with a retrieval cue. If this retrieval cue is later paired with other memory contents, it may be difficult to retrieve the old association and thus the old memory trace during retrieval (Anderson & Neely, 1966). We also know these effects from list learning: If subjects are asked to memorize a list with item pairings (e.g. “house” – “stone”) and later a second list that uses the same recall cues but new items (e.g. “house” – “door”), then learning the second list impairs recall of the first list. Other work, however, also pointed to the role of inhibition in forgetting to explain the results. This work has also shown that, contrary to popular belief, remembering can lead to forgetting (Anderson & Levy, 2007). The phenomenon of “retrieval induced forgetting” refers to a deterioration in memory performance for items that have not been relearned when their competitors have been learned. For this purpose, subjects first learn item pairs of a semantic category. They then have to recall some of the items in a category (“retrieval practice”), while others are not practised. It turns out that in a later retrieval phase, unpractised items of a practised category are more impaired than unpractised items of a non-practised category. This finding is interpreted by inhibitory processes between the items of a category – the retrieval of competitors leads to an inhibition of non-practised items in the retrieval practice phase. These results show that forgetting (to a non-pathological extent) is not so much an insufficient memory function, but may be a consequence of efficient information processing and storage.
1 Note that the results described here could later also be fully explained and re-evaluated by reconsolidation processes. The reconsolidation approach assumes that when a memory trace is recalled (here: questions about the pictures), the old memory trace is transformed into a retrievable, fragile state in which information can be changed or added to the existing trace (Lee et al., 2017).